Papers by Gaetano Scarano (UNIROMA1)
移動フレーム領域テクスチャ解析に基づくNEVI医用画像におけるアーチファクト除去【Powered by NICT】
IEEE Conference Proceedings, 2016
Performance analysis of blind channel estimation using second order statistics
In this contribution, we derive the performance of blind channel estimation without relying on pr... more In this contribution, we derive the performance of blind channel estimation without relying on prior knowledge of the symbol statistic. In particular, a lower bound on the performance of the maximum likelihood (ML) estimator of FIR channels is derived and expressed in a simple closed-form. The dependence of system parameters such as the roll-off of the raised cosine shaping filter, the channel spectrum, the SNR and the sample size is made explicit and an application example referring to a two ray communication channel is discussed. Simulation results are also provided to assess the validity of the proposed theoretical expressions in comparison with common approaches encountered in the literature.
Journal of Communications and Networks, Feb 1, 2022
In this paper we propose a novel client-transparent Dynamic Adaptive Streaming over HTTP (DASH)-a... more In this paper we propose a novel client-transparent Dynamic Adaptive Streaming over HTTP (DASH)-aware bandwidth allocation strategy. The approach, while being application layer transparent, guarantees fluidity to all users but it provides priority-based services to premium users, identified by their Willingness-To-Pay (WTP) profiles. Since different service qualities can be accommodated using WTP, the approach can be extended to XR services, immersive videos, live uplink streaming. To achieve this goal, the allocation problem is firstly formulated as a classical Game Theory problem whose closed form solution is clearly understood in the literature. In a nutshell, the WTPbased, Game Theoretically Optimal Bandwidth Allocation (WTP-GTOBA), firstly satisfies the minimum bandwidth needs of each user and then fairly distributes the residual bandwidth. WTP-GTOBA can be approximately implemented by a greedy, application layer transparent algorithm to be implemented at a DASHaware network element managing different neighbouring radio access stations. Thereby, the proposed method is suitable for integration of WTP and resource management among multiple service providers and heterogeneous client groups. Numerical simulations carried out under a realistic scenario show that the proposed approach outperforms state-of-the-art applicationlayer transparent competitors, providing premium quality and/or guaranteed fluidity to different users based on their WTP.
arXiv (Cornell University), Dec 5, 2019
The extraction of brain functioning features is a crucial step in the definition of brain-compute... more The extraction of brain functioning features is a crucial step in the definition of brain-computer interfaces (BCIs). In the last decade, functional connectivity (FC) estimators have been increasingly explored based on their ability to capture synchronization between multivariate brain signals. However, the underlying neurophysiological mechanisms and the extent to which they can improve performance in BCI-related tasks, is still poorly understood. To address this gap in knowledge, we considered a group of 20 healthy subjects during an EEG-based hand motor imagery (MI) task. We studied two well-established FC estimators, i.e. spectraland imaginary-coherence, and investigated how they were modulated by the MI task. We characterized the resulting FC networks by extracting the strength of connectivity of each EEG sensor and compared the discriminant power with respect to standard power spectrum features. At the group level, results showed that while spectral-coherence based network features were increasing in the controlateral motor area, those based on imaginary-coherence were decreasing. We demonstrated that this opposite, but complementary, behavior was respectively determined by the increase in amplitude and phase synchronization between the brain signals. At the individual level, we proved that including these network connectivity features in the classification of MI mental states led to an overall improvement in accuracy. Taken together, our results provide fresh insights into the oscillatory mechanisms subserving brain network changes during MI and offer new perspectives to improve BCI performance.
カメラ通信ぼけ除去:柔軟なフィルタ支援設計を用いたセミブラインド空間分数間隔適応等化器【Powered by NICT】
IEEE Conference Proceedings, 2016
Blind carrier frequency offset estimation for cross QAM constellations
ABSTRACT
MUSE: MUlti-lead Sub-beat ECG for remote AI based atrial fibrillation detection
Journal of Network and Computer Applications, Mar 1, 2023
IEEE Access, 2021
This paper tackles the problem of predicting the protein-protein interactions that arise in all l... more This paper tackles the problem of predicting the protein-protein interactions that arise in all living systems. Inference of protein-protein interactions is of paramount importance for understanding fundamental biological phenomena, including cross-species protein-protein interactions, such as those causing the 2020-21 pandemic of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Furthermore, it is relevant also for applications such as drug repurposing, where a known authorized drug is applied to novel diseases. On the other hand, a large fraction of existing protein interactions are not known, and their experimental measurement is resource consuming. To this purpose, we adopt a Graph Signal Processing based approach modeling the protein-protein interaction (PPI) network (a.k.a. the interactome) as a graph and some connectivity related node features as a signal on the graph. We then leverage the signal on graph features to infer links between graph nodes, corresponding to interactions between proteins. Specifically, we develop a Markovian model of the signal on graph that enables the representation of connectivity properties of the nodes, and exploit it to derive an algorithm to infer the graph edges. Performance assessment by several metrics recognized in the literature proves that the proposed approach, named GRAph signal processing Based PPI prediction (GRABP), effectively captures underlying biologically grounded properties of the PPI network.
Computationally efficient Maximum Likelihood phase acquisition for QAM constellations
2009 IEEE/SP 15th Workshop on Statistical Signal Processing, Aug 1, 2009
ABSTRACT
Evaluation of the Menzies method potential for automatic dermoscopic image analysis
In this work we present a Bayesian interpolation procedure to perform depth map upsampling. The d... more In this work we present a Bayesian interpolation procedure to perform depth map upsampling. The depth map prior is designed via an edge driven Markov Random Field. The upsampling procedure is computationally efficient and outperforms selected state of the art upsampling procedure; moreover it allows to perform depth map upsampling even without the reference high resolution luminance map.
IEEE Signal Processing Letters, 2020
Real-world networks are typically described in terms of nodes, links, and communities, having sig... more Real-world networks are typically described in terms of nodes, links, and communities, having signal values often associated with them. The aim of this paper is to introduce a novel Compound Markov random field model (Compound MRF, or CMRF) for signals defined over graphs, encompassing jointly signal values at nodes, edge weights, and community labels. The proposed CMRF generalizes Markovian models previously proposed in the literature, since it accounts for different kinds of interactions between communities and signal smoothness constraints. Finally, the proposed approach is applied to (joint) graph learning and signal recovery. Numerical results on synthetic and real data illustrate the competitive performance of our method with respect to other state-of-the-art approaches.
arXiv (Cornell University), Dec 21, 2020
Functional connectivity (FC) can be represented as a network, and is frequently used to better un... more Functional connectivity (FC) can be represented as a network, and is frequently used to better understand the neural underpinnings of complex tasks such as motor imagery (MI) detection in brain-computer interfaces (BCIs). However, errors in the estimation of connectivity can affect the detection performances. In this work, we address the problem of denoising common connectivity estimates to improve the detectability of different connectivity states. Specifically, we propose a denoising algorithm that acts on the network graph Laplacian, which leverages recent graph signal processing results. Further, we derive a novel formulation of the Jensen divergence for the denoised Laplacian under different states. Numerical simulations on synthetic data show that the denoising method improves the Jensen divergence of connectivity patterns corresponding to different task conditions. Furthermore, we apply the Laplacian denoising technique to brain networks estimated from real EEG data recorded during MI-BCI experiments. Using our novel formulation of the J-divergence, we are able to quantify the distance between the FC networks in the motor imagery and resting states, as well as to understand the contribution of each Laplacian variable to the total J-divergence between two states. Experimental results on real MI-BCI EEG data demonstrate that the Laplacian denoising improves the separation of motor imagery and resting mental states, and shortens the time interval required for connectivity estimation. We conclude that the approach shows promise for the robust detection of connectivity states while being appealing for implementation in real-time BCI applications.
Atrial Fibrillation Detection by Multi-Lead ECG Processing at the Edge
2021 IEEE Globecom Workshops (GC Wkshps), Dec 1, 2021
Atrial fibrillation is one of the most common arrhythmia events potentially causing heart failure... more Atrial fibrillation is one of the most common arrhythmia events potentially causing heart failures and thrombosis. Recently, many healthcare applications have been developed with the aim to provide a reliable real-time detection of such abnormal heartbeat behavior. The largest part of current solutions considers signal processing applied to electrocardiographic (ECG) segments recorded with wearable devices/sensors and specifically tailored to the number of available ECG leads. Differently, in this contribution, we present a lightweight machine learning algorithm for the analysis of ECG signals and atrial fibrillation detection, easily adaptable for both single and multi-lead architectures. Furthermore, we describe how the proposed scheme can be implemented with an edge computing approach that paves the way toward smart healthcare at home and remotely in general. In such framework, signal processing is not necessarily performed on a unique device, but it is conveniently split so as to let fast operations at the network edge, with storage and heavy computing being instead handled at cloud server side.
Deep Ll-PCA of Time-Variant Data with Application to Brain Connectivity Measurements
Ll-Principal Component Analysis (LI-PCA) is a powerful computational tool to identify relevant co... more Ll-Principal Component Analysis (LI-PCA) is a powerful computational tool to identify relevant components in data affected by noise, outliers, partial disruption and so on. Relevant efforts have been made to adapt its powerful summarization capacity to time variant data, e.g. in tracking the evolution of LI-PCA components. Here, we analyze a layered version of LI-PCA, to which we refer to as Deep LI-PCA. Deep LI-PCA is obtained by recursive application of two stages: estimation of LI-PCA basis and extraction of the first rank projector. The Deep LI-PCA is applied to repeated EEG connectivity measures and it proves relevant for identifying outliers, changes, and stable components. Moreover, at each layer, an in-depth analysis of the mean square error between the data applied at the input layer and the output projector is provided. The Deep LI-PCA allows to cope with outliers of different temporal extent as well as to extract the relevant common component at a reduced computational cost.
Comments on “Invariance property of Gaussian signals: A new interpretation, extension and applications”
Circuits Systems and Signal Processing, 1999
[1] L. Cheded, Invariance property of Gaussian signals: A new interpretation, extension and appli... more [1] L. Cheded, Invariance property of Gaussian signals: A new interpretation, extension and applications, Circuits, Systems, Signal Processing, vol. 16, no. 5, 1997. [2] G. Scarano, Cumulant series expansion of hybrid non-linear moment of complex random variables, 1EEL Transactions on Signal Processing, vol. 39, no. 4, April 1991. [3] G. Scarano, D. Caggiati, G. Jacovitti, Cumulant series expansion of hybrid non-linear moment o fn variates, IEEE Transactions on Signal Processing, vol. 41, no. I, January 1993.
EEG as Signal on Graph: a Multilayer Network model for BCI applications
2022 30th European Signal Processing Conference (EUSIPCO), Aug 29, 2022
Visually relevant point clouds features extraction by Radial Angular Point cloud filtering
2022 10th European Workshop on Visual Information Processing (EUVIP)
Invariance property of Gaussian signals: A new interpretation, extension and applications
Circuits Systems and Signal Processing, 1997
ABSTRACT
On the Underwater Acoustic Channel Effects on Uplink Multiple Access Techniques
OCEANS 2021: San Diego – Porto, Sep 20, 2021
Acoustic communication represents the most reliable technology to transmit data over long range u... more Acoustic communication represents the most reliable technology to transmit data over long range underwater links, but unfortunately data rates performance suffers from the poor available bandwidth, the low speed of sound and the impairments caused by the channel. Such limitations impact not only on physical layer issues, but also when dealing with medium access control. In fact, the access opportunities for users deployed in a cellular-like underwater network are strongly influenced by the channel, as delay spread variability and Doppler can worsen the performance not only in terms of intersymbol interference but also in terms of multi-user interference. In this regard, the following contribution is aimed to investigate the role played by the channel in terms of access opportunities for time, frequency, code - division multiple access, that are the best known channelization based techniques, as well as for random access as ALOHA.
Zenodo (CERN European Organization for Nuclear Research), Jan 25, 2018
The boost of signal processing on graph has recently solicited research on the problem of identif... more The boost of signal processing on graph has recently solicited research on the problem of identifying (learning) the graph underlying the observed signal values according to given criteria, such as graph smoothness or graph sparsity. This paper proposes a procedure for learning the adjacency matrix of a graph providing support to a set of irregularly sampled image values. Our approach to the graph adjacency matrix learning takes into account both the image luminance and the spatial samples' distances, and leads to a flexible and computationally light parametric procedure. We show that, under mild conditions, the proposed procedure identifies a near optimal graph for Markovian fields; specifically, the links identified by the learning procedure minimize the potential energy of the Markov random field for the signal samples under concern. We also show, by numerical simulations, that the learned adjacency matrix leads to a higly compact spectral wavelet graph transform of the so obtained signal on graph and favourably compares to stateof-the-art graph learning procedures, definetly matching the intrinsic signal structure.
Uploads
Papers by Gaetano Scarano (UNIROMA1)